[ BDSN ] 6 AEO myths you can ignore for your brand

STUMP BEZOS

TikTok snd Youtube are increasingly carrying the growth of social commerce. How much of ALL e-commerce sales is social commerce expected to be in 2027?

[ Answer at bottom of email ]

💰 AMAZON ALREADY TOLD YOU WHICH IMAGE SHOULD BE #2

The average Prime shopper converts 74% of the time. They came to buy and they will buy something. The only question is whose listing gets the sale, and that gets decided by whether your images check the boxes on their mental list in the order they care about.

Anthony Cofrancesco ran a BDSC session on exactly this. Eight years selling on Amazon, formerly at PickFu and Data Dive, and his whole method comes down to pattern recognition. No camera. No Photoshop. Just knowing what buyers want to see and putting it in front of them first. The part you can do today takes about 15 minutes.

Step one: let Amazon tell you what matters

Open Seller Central. Go to Growth, then Product Opportunity Explorer. Enter your main keyword and find the Key Success Attributes section. Amazon lists the three to five things shoppers in that niche care about most, built from its own search and purchase data.

For dish towels, the list came back as material quality first (100% cotton, weave pattern), sizing second (kitchen towel vs. dish cloth), multi-pack options third.

Now open your listing. Does image two cover attribute one? Does image three cover attribute two? For most listings the answer is no. The sizing graphic sits in slot six. The material shot doesn't exist. Image two is a lifestyle photo that answers nothing.

If the order doesn't match, reorder. Move the sizing graphic from six to three. Change nothing else. Watch unit session percentage for 14 days against your prior 30-day average. This is the lowest-risk test on Amazon. You're rearranging what you already own.

Step two: break up the kitchen sink graphic

Find your busiest image. The one that says soft, absorbent, skin-friendly and durable all at once, with four icons and a wall of text. Shoppers are distracted. Dog barking, kid on the counter, phone in one hand. Dense graphics don't get read.

Pick the single most important point in that graphic (ideally attribute one from your POE list) and make it the only thing on the image. Anthony's prompt for ChatGPT or Gemini:

 "Create a single feature graphic for [product] that focuses ONLY on [one specific feature]. Show [product] on the left, feature callout on the right with large text and a two-sentence description. Clean, minimal design with lots of white space."

Swap the cluttered graphic for the focused one. Test it the same way. One change, 14 days.

Step three: add the type you're missing

There are only 18 kinds of images on Amazon. Anatomy graphics with arrows and callouts. Sizing graphics with a coffee mug for scale. Application grids showing four uses. Comparison charts. Testimonial graphics. How-to steps. Metrics. Before and after.

Screenshot the secondary images of your top five to ten competitors and tag each one by type. Any type that shows up in 40% or more of the niche and is missing from your stack is a hole. The most common gaps: no anatomy graphic, no sizing graphic, no testimonial graphic.

Testimonial is the cheapest fix. Pull three or four of your best reviews and prompt:

"Create a testimonial graphic showing three customer reviews. Each with a 5-star rating visual, a one to two sentence quote, and a first name. Clean layout, [brand color] accents, white background, VERIFIED CUSTOMERS header at top."

Thirty minutes, done.

The rules that keep this safe

Risk equals action times time. Change one thing. Check every two or three days. Revert if conversion drops more than 10% and holds for a week. Keep anything that improves, even 2%.

Then do it again next month. One test a month, twelve a year. Most sellers don't touch their images for a year, which is why a reorder that costs nothing can pass them.

Anthony's full walkthrough, the 18-type checklist, AI prompts for every image type, and his 12-month testing calendar are all in the replay inside the Billion Dollar Sellers Club, next to a growing library of sessions like it. Try it for $9 at billiondollarsellers.club.

🛠️ BDSN SOFTWARE TOOL of the DAY 🛠️

Titan AI is an AI agent for Amazon sellers that works toward your plan instead of waiting to be asked.

This first look shows the next Titan AI release, out next week.

Built for 7 and 8 figure Amazon and e-commerce brand owners.

You set your goal in one sentence and Titan AI drafts it from your own numbers. Headquarters puts your whole business in one view, with the actions waiting for your approval.

Tasks and Workflows run your routine Amazon PPC work on a schedule, like a weekly search term review that finds 3 terms spending without an order, shows you what they cost, and negates them once you approve.

Try Titan AI on your own Amazon account with a 14 day Titan Guest Pass.

🚀 BEFORE YOU PAY FOR AEO SERVICES, READ THIS

Google’s Official Word on AI Search: It’s Still SEO

Google published its official guide to showing up in AI Overviews and AI Mode (last updated July 2026). The short version: optimizing for AI search is optimizing for search.

Why it’s still SEO

AI Overviews and AI Mode pull from Google’s normal index and ranking systems.
They retrieve pages through core ranking, then write an answer with links to those pages.
They also run “query fan-out”: the model fires off related searches to gather more info, so one question becomes several.
A page must be indexed and eligible for a snippet to appear. Indexing is never guaranteed.
On AEO and GEO, Google’s position is that it’s all just SEO.

Myths Google says you can ignore

llms.txt and special AI files: Google Search doesn’t use them. They won’t help or hurt.
Chunking content into tiny pieces: Not needed. Google can pull the relevant part from a multi-topic page. There’s no ideal page length.
Rewriting content just for AI: The systems understand synonyms and meaning. You don’t need every long-tail variation.
Chasing inauthentic mentions: Less helpful than it looks. Both quality and spam systems feed the AI features.
Obsessing over structured data: Not required, and there’s no special AI schema. Still useful for rich results.
Writing a page for every fan-out query: Doing this to manipulate AI answers can violate Google’s scaled content abuse policy. It also doesn’t work long term.

Red flags when buying AEO services

If the pitch is llms.txt, “AI-ready” chunking, special schema, or mention seeding, they’re selling things Google says don’t matter.
Google says no third-party tool has access to its internal ranking or AI systems. Be wary of any tool claiming “internal Google metrics.”
Ask one question: what are you doing that isn’t normal SEO?
Google has a separate guide on evaluating third-party SEO advice. Worth sending to your team before you sign anything.

What actually moves the needle

Non-commodity content. First-hand experience, original data, a real point of view. Generic listicles any AI could write are the weak spot. Google says this will likely matter more than anything else in the guide.
Clean technical basics. Crawlable, indexed, fast, no duplicate mess. Perfect code isn’t required.
Good images and video. They give you more ways to appear.
Merchant Center feeds and Google Business Profile. These get products and local businesses into AI responses.
Measurement. Search Console now has a Generative AI performance report.
Agent readiness. Browser agents are already visiting sites, and commerce protocols like UCP are emerging. Worth watching.

Bottom line: Google also notes plenty of content ranks fine with no overt SEO at all. Make something worth citing, keep your site clean, and skip the hacks.

Source: Optimizing your website for generative AI features on Google Search

Which of your Black Friday deals will make money?

You set your Black Friday deals weeks ago, when the numbers looked good on revenue.

Then Black Friday week starts and the sales dashboard looks great. The deal fees and ad spend land later, in the settlement. That's when many sellers find out a few of those deals lost money on every unit.

DataDoe shows you that before Black Friday, while you can still act on it. It connects Seller Central and Amazon Ads to Claude or whatever AI you already use.

Ask which deals still make money at the deal price, after COGS, fees and ads, and you get the answer SKU by SKU.

Then fix what you can: pause the ads behind the losing deals and reprice the products that aren't in one. Every change waits for your OK before it reaches Amazon.

Setup takes about five minutes. You get a 14-day free trial, then it's $97 a month. Code KEVINKING50 at checkout takes 50% off the first month.

SEE WHICH DEALS MAKE MONEY →

📺 YOU GOTTA SEE THIS

If you are generating millions of dollars a year on Amazon but ignoring your email list, you are likely leaving 30% to 40% of your gross revenue on the table.

In this episode of Marketing Misfits, hosts Norm Farrar and Kevin King break down the exact strategies top sellers use to turn email into an absolute cash-printing machine.

They expose why focusing on "Open Rates" is a massive mistake (thanks in part to Apple's privacy updates), why "pretty" HTML emails can hurt deliverability, and how using link shorteners like Bitly will get you instantly blocked by spam filters.

Kevin and Norm also reveal the behind-the-scenes AI system they use at their agency, DragonFish, which builds 16 specialized AI "brains" to craft hyper-personalized emails that hit inboxes at the exact moment a customer is checking them.

If you think email marketing is "dead" or "spammy," this episode will completely change your perspective—and your profit margins.